Introduction to Mobile Robotics Welcome Wolfram Burgard 1 Today - - PowerPoint PPT Presentation

▶
introduction to mobile robotics welcome
SMART_READER_LITE
LIVE PREVIEW

Introduction to Mobile Robotics Welcome Wolfram Burgard 1 Today - - PowerPoint PPT Presentation

Oct 03, 2022 415 likes •1.05k views

Introduction to Mobile Robotics Welcome Wolfram Burgard 1 Today This course Robotics in the past and today 2 Organization Self Study lecture recordings, no on-site lectures Thursday 13:00 14:00 homework, practical

slide-1
SLIDE 1

1

Wolfram Burgard

Welcome Introduction to Mobile Robotics

slide-2
SLIDE 2

Today

  • This course
  • Robotics in the past and today

2

slide-3
SLIDE 3

3

Organization

  • Self Study

lecture recordings, no on-site lectures

  • Thursday

13:00 – 14:00

homework, practical exercises (Python), discussions

  • Web page:

www.informatik.uni-freiburg.de/~ais/

  • Exam: Oral or written
slide-4
SLIDE 4

People

Teaching:

  • Wolfram Burgard

Teaching assistants:

  • Daniel Büscher
  • Marina Kollmitz
  • Lukas Luft

4

slide-5
SLIDE 5

5

Goal of this course

  • Provide an overview of problems and

approaches in mobile robotics

  • Probabilistic reasoning: Dealing with noisy

data

  • Hands-on experience
slide-6
SLIDE 6

Content of this Course

  • 1. Linear Algebra
  • 2. Wheeled Locomotion
  • 3. Sensors
  • 4. Probabilities and Bayes
  • 5. Probabilistic Motion Models
  • 6. Probabilistic Sensor Models
  • 7. Mapping with Known Poses
  • 8. The Kalman Filter
  • 9. The Extended Kalman Filter

10.Discrete Filters 11.The Particle Filter, MCL

6

  • 12. SLAM: Simultaneous Localization

and Mapping

  • 13. SLAM: Landmark-based

FastSLAM

  • 14. SLAM: Grid-based FastSLAM
  • 15. SLAM: Graph-based SLAM
  • 16. Techniques for 3D Mapping
  • 17. Iterative Closest Points

Algorithm

  • 18. Path Planning and Collision

Avoidance

  • 19. Multi-Robot Exploration
  • 20. Information-Driven Exploration
  • 21. Summary
slide-7
SLIDE 7

Reference Book

Thrun, Burgard, and Fox: “Probabilistic Robotics”

slide-8
SLIDE 8

Relevant other Courses

  • Foundations of Artificial Intelligence
  • Computer Vision
  • Machine Learning
  • and many others from the area of cognitive

technical systems.

8

slide-9
SLIDE 9

Opportunities

  • Projects
  • Practicals
  • Seminars
  • Thesis
  • … your future!

9

slide-10
SLIDE 10

Autonomous Robot Systems

  • perceive their environment and
  • generate actions to achieve their goals.

environment sense act

model

slide-11
SLIDE 11

Tasks Addressed that Need to be Solved by Robots

  • Navigation
  • Perception
  • Learning
  • Cooperation
  • Acting
  • Interaction
  • Robot development
  • Manipulation
  • Grasping
  • Planning
  • Reasoning

…

slide-12
SLIDE 12

13

Robotics Yesterday

slide-13
SLIDE 13

14

Current Trends in Robotics

Robots are moving away from factory floors to

  • Entertainment, toys
  • Personal services
  • Medical, surgery
  • Industrial automation

(mining, harvesting, …)

  • Hazardous environments

(space, underwater)

slide-14
SLIDE 14

15

Shakey the Robot (1966)

slide-15
SLIDE 15

16

Shakey the Robot (1966)

slide-16
SLIDE 16

Robotics Today

  • Lawn mowers
  • Vacuum cleaners
  • Self-driving cars
  • Logistics
  • …

18

slide-17
SLIDE 17

19

The Helpmate System

slide-18
SLIDE 18

Autonomous Vacuum Cleaners

slide-19
SLIDE 19

Autonomous Lawn Mowers

21

slide-20
SLIDE 20

22

DARPA Grand Challenge

[Courtesy by Sebastian Thrun]

slide-21
SLIDE 21

Walking Robots

[Courtesy by Boston Dynamics]

slide-22
SLIDE 22

Androids

Overcoming the uncanny valley

[Courtesy by Hiroshi Ishiguro]

slide-23
SLIDE 23

Driving in the Google Car

slide-24
SLIDE 24

Autonomous Motorcycles

[Courtesy by Anthony Levandowski]

slide-25
SLIDE 25

The Google Self Driving Car

29

slide-26
SLIDE 26

Folding Towels

slide-27
SLIDE 27

31

Rhino

(Univ. Bonn + CMU, 1997)

slide-28
SLIDE 28

32

Minerva

(CMU + Univ. Bonn, 1998)

Minerva

slide-29
SLIDE 29

Robotics in Freiburg

33

slide-30
SLIDE 30

Autonomous Parking

slide-31
SLIDE 31

Autonomous Quadrotor Navigation

Custom-built system: laser range finder inertial measurement unit embedded CPU laser mirror

slide-32
SLIDE 32

Precise Localization and Positioning for Mobile Robots

slide-33
SLIDE 33

Obelix – A Robot Traveling to Downtown Freiburg

slide-34
SLIDE 34

The Obelix Challenge (Aug 21, 2012)

slide-35
SLIDE 35

The Tagesthemen-Report

slide-36
SLIDE 36

Brain-controlled Robots

40

slide-37
SLIDE 37

Teaching: Student Project on the Autonomous Portrait Robot

slide-38
SLIDE 38

Final Result

slide-39
SLIDE 39

Other Cool Stuff from AIS

43

slide-40
SLIDE 40

Accurate Localization

  • KUKA omniMove

(11t)

  • Safety scanners
  • Error in the area of

millimeters

  • Even in dynamic

environments

slide-41
SLIDE 41
  • Fuselage assembly
  • 20 vehicles to transport industrial robots for

drilling and filling of 60,000 fasteners in

  • 6 vehicles for logistics of parts, work stands

and fuselages

26 Units installed at Boeing

slide-42
SLIDE 42
slide-43
SLIDE 43

Deep Learning to Manipulate from Parallel Interaction

Source: Google Research Blog

slide-44
SLIDE 44

Learning User Preferences

  • Task preferences are subjective
  • Fixed rules do not match all users
  • Constantly querying humans is suboptimal
  • How to handle new objects?

Where does this go?

slide-45
SLIDE 45

?

Collaborative Filtering

  • …
slide-46
SLIDE 46

Collaborative Filtering

  • …
slide-47
SLIDE 47

Online Prediction of Preferences

slide-48
SLIDE 48

Localization in Urban Environments

  • Inaccurate (if even available) GPS signal
  • No map
  • Limited Internet
slide-49
SLIDE 49

Motivation

slide-50
SLIDE 50

Example

slide-51
SLIDE 51

Example contin.

Text: irpostbankfmarzcenter tllgi Matched Landmarks:

  • Postbank finanzcenter

Text: melange Matched Landmarks:

  • Melange
  • Melange

Text: casanova Matched Landmarks:

  • Casanova
slide-52
SLIDE 52

Example

slide-53
SLIDE 53

Deep Learning Applications

  • RGB-D
  • bject

recognition

  • Images

human part segmentation

  • Sound

terrain classification

slide-54
SLIDE 54
  • Fusion layers automatically learn to combine

feature responses of the two network streams

  • During training, weights in first layers stay fixed

DCN for Object Recognition

slide-55
SLIDE 55

Learning Results

Method RGB Depth RGB-D CNN-RNN 80.8 78.9 86.8 HMP 82.4 81.2 87.5 CaRFs N/A N/A 88.1 CNN Features 83.1 N/A 89.4 This work, Fus-CNN 84.1 83.8 91.3

  • Category-Level Recognition [%] (51 categories)
  • [Lai et. al, 2011]
slide-56
SLIDE 56

Network Architecture

  • Fully convolutional network
  • Contraction and expansion of network input
  • Up-convolution operation for expansion
  • Pixel input, pixel output
slide-57
SLIDE 57

Deep Learning for Body Part Segmentation

  • Input Image
  • Ground Truth
  • Segmentation

mask

slide-58
SLIDE 58

Deep Learning for Terrain Classification using Sound

slide-59
SLIDE 59

Network Architecture

  • Novel architecture designed for

unstructured sound data

  • Global pooling gathers statistics of learned

features across time

slide-60
SLIDE 60

Data Collection

Asphal t Wood Offroa d Cobble Stone Paving Grass Mowe d Grass Carpet Linoleu m P3-DX

slide-61
SLIDE 61

Results - Baseline Comparison

16.9% improvement over the previous state of the art 99.41% using a 500ms window

(300ms window) [1] [2] [3] [4] [5] [6]

[1] T. Giannakopoulos, K. Dimitrios, A. Andreas, and T. Sergios, SETN 2006 [2] M. C. Wellman, N. Srour, and D. B. Hillis, SPIE 1997. [3] J. Libby and A. Stentz, ICRA 2012 [4] D. Ellis, ISMIR 2007 [5] G. Tzanetakis and P. Cook, IEEE TASLP 2002 [6] V. Brijesh , and M. Blumenstein, Pattern Recognition Technologies and Applications 2008

slide-62
SLIDE 62

Thank you

… and enjoy the course!

67